|
(11) | EP 1 316 480 B1 |
| (12) | EUROPEAN PATENT SPECIFICATION |
|
|
| (54) |
Pre-Crash Threat Assessment System in a Remote Crash detection System Precrash Bedrohungbewertungssystem in einem Crashdetektionssytem Système dévaluation de la sévérité d'un évênement avant collision dans un système de détection à distance d'accident |
|
|
|||||||||||||||||||||||||||||||
| Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention). |
[0001] The present invention relates generally to crash detection systems for automotive vehicles and more particularly to a pre-crash threat assessment system for a crash detection system.
[0002] Due to the current density of traffic on the roads, motor vehicle operators are flooded with information and consequently operating a motor vehicle is a complex procedure in which various situations arise where the operator has limited, little, or no time to react or to manually engage safety measures.
[0003] Many previously known crash detection systems have incorporated crash detection algorithms based on sensed data. The application of remote sensing systems using radar, lidar, and vision based technologies for object detection, tracking, alarm processing, and potential safety countermeasure deployment is well known in the art. These can range in their function from those which do not give accurate enough predictions to those which are extremely complex and therefore expensive and requiring significant processing power. The latter type includes that disclosed in US 6085151A.
[0004] Based on range and bearing information provided by radar, lidar, or vision based systems and based on additional information obtained from the host vehicle sensors, various algorithms have been used to track the path of a host vehicle, to track the path of a target, and to estimate the future position of objects in the host vehicle path.
[0005] Safety systems, such as airbags and motorized safety belt pre-tensioners, activate after contact occurs between two vehicles. A typical accident occurs within 90ms, whereas a typical airbag deploys within approximately 70ms. Time minimization between the start of an accident and the start of safety system deployment is therefore crucial.
[0007] Currently, accident prediction algorithms are employed primarily for accident warning/avoidance and operate typically within a range larger than 30 meters between host and target vehicles. In the event that a collision is unavoidable, however, the range is less than, and often considerably less than, 30 meters. Therefore, damage minimization techniques must predict an unavoidable collision and deploy safety measures within a short time.
[0008] The limitations associated with current accident damage minimization techniques have made it apparent that a new technique to minimize damage to a vehicle or vehicle operator is needed. The new technique should predict a target vehicle position with respect to a host vehicle and should also substantially minimize the computational time for threat assessment calculations. The present invention is directed to these ends.
[0009] It is an object of this invention to provide an improved pre-crash sensing system for a motor vehicle.
[0011] Preferably the system further includes a second safety device actuator, coupled to the host object, for activating a second safety device, and wherein the safety device controller determines a second collision potential assessment based on the future transverse position of the target object relative to a second threshold fraction of the host object width and controls the second safety device actuator in response to the second collision potential assessment.
[0014] Preferably the first safety device is at least one of a pre-arming device for an airbag and a motorized safety belt pre-tensioner.
[0015] Advantageously the remote sensor is one of a lidar sensor, a radar sensor system, a vision system and a combination of radar, lidar and vision systems.
[0016] The first safety device actuator may respond to the first threshold fraction of the host object width in combination with safety device specific deployment criteria.
[0018] Preferably the step of sensing further comprises the step of sensing multiple target objects in the near zone of the host object.
[0019] Advantageously the step of determining further comprises the step of determining whether the potential for collision of the host object and the target object is within a second threshold fraction of the host object width.
[0020] The invention will now be described by way of example with reference to the accompanying drawing of which:-
FIGURE 1 is a pre-crash assessment system in accordance with a preferred embodiment of the present invention;
FIGURE 2 is a block diagram of a remote sensing based pre-crash threat assessment system in accordance with a preferred embodiment of the present invention;
FIGURE 3 is an exemplary illustration of a pre-crash scenario in accordance with a preferred embodiment of the present invention; and
FIGURE 4 is a block diagram of a pre-crash threat assessment and safety device activation system in accordance with a preferred embodiment of the present invention.
[0021] The present invention is illustrated with respect to a pre-crash threat assessment and safety device activation system, particularly suited to the automotive field. The present invention is, however, applicable to various other uses that may require pre-crash threat assessment, as will be understood by one skilled in the art.
[0022] Referring to Figure 1, a pre-crash assessment system 1, including a first target object here illustrated as a first target vehicle 2 imminently colliding with a host vehicle 3, is illustrated.
[0023] The pre-crash assessment system includes a high frequency remote sensor (or remote sensing system, as will be understood by one skilled in the art) 4 coupled to the host vehicle 3. The sensor 4 detects vehicle states of the first target vehicle 2. Examples of the vehicle states are position and velocity.
[0024] A first safety device actuator 5 is coupled to the host vehicle 3 and pre-arms a first safety device 6, here embodied as a pre-arming device for an airbag. A second safety device actuator 7 is also coupled to the host vehicle 3. The second safety device actuator 7 activates a second safety device 8, here embodied as a motorized safety belt pre-tensioner. It is important to note is that numerous actuators and safety devices may be added to the system as needed by the manufacturer.
[0025] A safety device controller 9 is also coupled to the host vehicle 3. From the remote sensing system detected target vehicle states, the safety device controller 9 calculates target vehicle dynamics with respect to the host vehicle. Examples of target vehicle dynamics are velocities and accelerations.
[0026] The safety device controller 9 generates a threshold assessment based on the sensed target vehicle dynamics and also generates a control signal based on the host vehicle dynamics. The sensed information, in conjunction with host vehicle dynamic information, is used by the controller for countermeasure activation decision making.
[0027] The controller 9 estimates whether a potential for crash between the host vehicle 3 and the first target vehicle 2 is within the first threshold for the first safety device actuator 5. The controller 9 also estimates whether a potential for crash between the host vehicle 3 and the first target vehicle 2 is within the second threshold for the second safety device actuator 7.
[0028] In the current embodiment the estimation is made through a pre-crash algorithm. The algorithm is defined by the comparison of the predicted position of the target object, when the front of the target vehicle is in line with the front of the host vehicle 3 (i.e., when it is in the OX plane of the coordinate system, which is illustrated at the front of the host vehicle 3), with a fraction of the host vehicle width (ideally half the width (W/2)) plus additional adjustable tolerance allowed by the host vehicle systems.
[0029] The fraction of width used is dependent on host vehicle parameters. The adjustable tolerance zone is a function of the last measured position of the target object, the target object direction of travel and the target object position when it comes in line with the front of the host vehicle.
[0030] The fraction of the width and the adjustable tolerance values are either pre-set into the system 12 during manufacture or later through a separate sensor that reads vehicle dimensions, which is subsequently attached to the host vehicle 3. The threat is severe, and, in the current embodiment, within the first actuator threshold when the target vehicle 2 predicted position is less than the fraction of width used. The threat is severe, and, in the current embodiment, within the second actuator threshold when the target vehicle predicted position is substantially less than the fraction of width used.
[0031] The safety device controller 9 further sends control signals to the host vehicle Controller Area Network Bus (CAN) 10, which controls the first safety device actuator 5 and the second safety device actuator 7 in response to threat assessment evaluations, as will be understood by one skilled in the art. The operations of the controller 9 will be discussed in detail later.
[0032] Referring to Figure 2, a block-diagram of the remote sensing based pre-crash threat assessment system 12 is illustrated. The current invention addresses only threat assessment aspects of the system 12 (for pre-crash sensing purposes) with radar, lidar, or vision sensor based remote sensing systems.
[0033] The system 12 starts when operation block 13, which engages signal processing and detection algorithms, receives radar sensor data and predetermined detection thresholds. The radar sensor data is generated when an object impedes the radar pulse and reflects the pulse back to the radar sensor on the host vehicle. The detection thresholds are pre-set based on vehicle specific acceptable probability of detection and false alarm rates. Subsequently, operation block 13 sends the data and noise accompanying the signal, as will be understood by one skilled in the art, to operation block 14. The probability of detection and false alarm rates have significant effects on track initiation and track quality.
[0034] Operation block 14 associates the data from operation block 13 and engages an object tracking algorithm. Operation block 14 then sends the calculated track estimates of the object and the tracking error estimate signals to operation block 16, as will be understood by one skilled in the art.
[0035] Host vehicle dynamic data is also sent to the operation block 16 from the host vehicle dynamic sensing systems. Using this combination of received data, operation block 16 estimates the future states of the host vehicle and target object and sends this data to operation block 18. An evaluation is then made in operation block 18 of the potential for collision between the host vehicle and the target object.
[0036] Operation blocks 16 and 18 are the threat assessment portion of the system 12, which will be discussed in detail later. Operation block 18 also evaluates whether the countermeasure activation criteria are met for various countermeasure actions under consideration. Subsequently, operation block 18 sends actuation signals to the Controller Area Network Bus (CAN) of the host vehicle, which engages the safety devices (countermeasures), as will be understood by one skilled in the art.
[0037] In the current invention, the host vehicle dynamics data is used for evaluating countermeasure deployment criteria only. Host vehicle dynamic data is not used to predict the future state of the object with respect to the host vehicle. Instead, the object future state, with respect to the host vehicle future state, is predicted through remote sensor obtained object position and bearing data (as a function of time).
[0038] The embodied approach-is further clarified in Figure 3, which illustrates an example of a pre-crash scenario 20. S1, S2, S3 and S4 are the positions of the host vehicle 22 at four consecutive times t1, t2, t3 and t4 respectively. T1, T2, T3 and T4 are the positions of a target object (ideally a target vehicle) 24 being tracked by the remote sensing system 12.
[0039] Multiple objects are typically tracked, but to explain the invention, only the object that has the highest potential for crash is shown.
[0040] R is the radial distance from the origin of the coordinate system, which is positioned at the front centreline of the host vehicle, to the nearest scattering centre 27 of the target object 24. θ is the angle made by the nearest scattering centre on the target object 24 with the Y-axis of the coordinate system. A radar sensor 26 is ideally attached to the front of the host vehicle 22.
[0041] The Y-axis of a coordinate system is configured along the front central line of the host vehicle 22. In the Cartesian coordinate system, with origin at 0 on the host vehicle 22:
X1= R1 sin (θ1) and Y1= R1 cos(θ1) at time t1,
X2= R2 sin (θ2) and Y2= R2 cos(θ2) at time t2,
X3= R3 sin (θ3) and Y3= R3 cos(θ3) at time t3
and
X4= R4 sin (θ4) and Y4= R4 cos(θ4) at time t4.
[0042] In this manner, the X and Y coordinates of the target object 24 with respect to the coordinate system, fixed to the host vehicle and moving therewith, is determined as a function of time, which is here represented as: t1, t2, t3 and t4.
[0043] By a numerical differentiation process, as will be understood by one skilled in the art, the X and Y components of the target object 24 velocities and accelerations are obtained.
[0044] When Doppler velocity measurement capable radar or lidar sensors are used, the relative velocity information can be directly measured. In this situation, only X and Y components of the acceleration data need to be calculated using numerical differentiation procedures.
[0045] From the Y component of current position (e.g. Y4 corresponding to T4) and Y components of velocity (e.g. Vy4) and acceleration (e.g. Ay4), with respect to the moving coordinate system OXY, the time (tp) required for the target object 24 to be line OX is calculated through the equation:
[0046] From the X component of the current position (e.g. X4) and the X component of velocity (e.g. Vx4) and acceleration (e.g. Ax4), with respect to the moving coordinate system OXY, and the time (tp) at which the target object 24 will be on line OX, the X position at time (tp) is predicted through the equation:
Xtp = X4+ VX4 * tp+ 0.5*AX4*(tp)2.
[0047] The severity of threat (confidence level) is then assessed by comparing this predicted Xtp position (X at time tp) with a fraction of the host vehicle width (ideally half the width (W/2)) plus additional variable tolerance allowed by the host vehicle systems.
[0048] The threat is severe within the first actuator threshold when the target vehicle predicted position (Xtp) is less than the fraction of width used and is severe within the second actuator threshold when the target vehicle predicted position (Xtp) is substantially less than the fraction of width used.
[0049] This algorithm employs only the remote sensing systembased measurements. This algorithm uses substantially less time to run than previous algorithms, which also considered the vehicle dynamics of the host vehicle in the evaluation of the crash assessment situations. Through this algorithm, the system substantially achieves the goal of providing a threat assessment means with short computational time requirements for threat assessment evaluations, thus providing more time for deployment of necessary countermeasures in a non-avoidable collision situation.
[0050] In Figure 4, in view of and referencing components illustrated in Figure 2 and Figure 3, a block diagram of the operation of a pre-crash threat assessment and safety device activation system 30, in accordance with one embodiment of the present invention, is illustrated. The logic starts in operation block 32 by obtaining the range (R) and the bearing (θ) of the target object 24 from the host vehicle remote sensor 26. Subsequently, operation block 34 activates and the object track record is updated from the data sent from operation block 32.
[0051] Operation block 36 then activates, and logic operative to estimate the velocity (Vy) and acceleration (AY) of the target object 24 is engaged. The data from operation block 34 is ideally filtered prior to estimation in operation block 36, as will be understood by one skilled in the art.
[0052] Operation block 38 then activates, which uses the longitudinal distance with respect to the 0X plane, and calculates the intercept time (tp) of the target object 24.
[0053] From this intercept time (tp), operation block 40 activates and logic is engaged, operative to predict the lateral position (X at time tp) of the target object 24, as was previously discussed.
[0054] Operation block 42 subsequently activates and assesses the potential for collision and evaluates the confidence levels associated with the position prediction. The confidence levels are the thresholds within which the safety device controller, incorporated in the host vehicle 22, activates a countermeasure in response to the position prediction and the meeting of device specific deployment criteria. The confidence levels are assessed, in the current embodiment, by comparing the predicted Xtp position (X at time tp) with the host vehicle half width (W/2) plus additional variable tolerance allowed by the host vehicle systems.
[0055] A decision is then made in inquiry block 44 as to whether controller confidence levels are met. For a negative response, operation block 32 again is performed and the block diagram 30 operates for a subsequent position, velocity and acceleration of the target object 24 with respect to the host vehicle 22.
[0056] Otherwise, the confidence levels are compared to the threshold of the first safety device (embodied as a pre-arming device for an airbag) and the threshold of the second safety device (embodied as a motorized belt pre-tensioner) in operation block 46. These threshold comparisons and safety device specific criteria are used to reach the countermeasure deployment decision.
[0057] A determination is then made in inquiry block 48 as to whether either safety device deployment criterion is met. Two safety device thresholds for two safety devices are included in the current embodiment, however numerous other safety device thresholds for numerous other safety devices may be added as needed for vehicle safety. In operation block 50, if either safety device deployment criterion is met, the controller sends a signal to the first safety device actuator or the second safety device actuator for the respective countermeasure (safety device) activation.
[0058] Otherwise, operation block 32 again activates and the block diagram 30 operates for a subsequent position, velocity and acceleration of the target object 24 with respect to the host vehicle 22.
[0059] The current embodiment combines this efficient approach for threat assessment with advanced techniques for radar based object tracking. The quality of threat assessment depends upon the quality of tracking. Advanced filtering and tracking techniques such as Alpha-Beta-Gamma filtering and adaptive Kalman filtering are used to initiate and maintain quality tracks for objects in the near zone of the host vehicle. These filtering and tracking techniques improve the reliability, robustness, and confidence levels of the threat assessment predictions without significantly sacrificing processing speeds, as will be understood by one skilled in the art.
[0060] In operation, in view of FIGURE 3, when a target object 24 comes in range of the radar sensor 26 of the host vehicle 22, logic operates to track the target object 24 and estimate the future position, velocity and acceleration of the target object 24 relative to the host vehicle 22. An assessment is then conducted for the confidence of collision and whether the collision will require pre-arming the airbag or activating the motorized belt pre-tensioners. Ideally, to minimize nuisance activation, the motorized belt pre-tensioner is activated when there is a higher certainty of collision. A lower threshold is used for pre-arming the airbag.
[0061] From the foregoing, it can be seen that there has been brought to the art a new remote sensing based pre-crash threat assessment system.
[0062] It is to be understood that the preceding description of the preferred embodiment is merely illustrative of some of the many specific embodiments that represent applications of the principles of the present invention. Numerous and other arrangements could be evident to those skilled in the art without departing from the scope of the invention, as defined by the set of claims.
a remote sensor (4,26) coupled to the host object for detecting an object dynamic of the target object (2,24) in relation to the host object,
a first safety device actuator (5), coupled to the host object, for activating a first safety device (6),
the system being characterised in that it further comprises
a pre-crash algorithm predicting the position of the target object within a coordinate system (oxy) defined by a transverse axis (ox) extending along the forward most edge and the longitudinal axis (oy) of the host object, based upon the target object dynamic,
the algorithm predicting the transverse predicted position (Xtp) of the target object from the origin at a predicted time (tp) when the target object crosses the transverse axis(ox),
a safety device controller (9) coupled to the host object for determining a collision potential assessment based on the transverse predicted position (xep) of the target object relative to a first threshold fraction of the host object width, and for controlling the first safety device actuator in response to the collision potential assessment.
sensing a target object (2, 24) in a near zone of the host object (3, 22),
tracking a current object dynamic the target object (2, 24) with respect to a coordinate defined system by a transverse axis extending along the forward most edge and the longitudinal axis of the host object,
predicting the transverse predicted positon of the target object from the origin at a predicted time when the target object crosses the transverse axis, based upon the target object dynamic,
determining a collision potential assessment by comparing the transverse predicted position of the target object to a threshold fraction of the host object width and
actuating a safety device in response to the determined collision potential assessment.
einen Ferngeber (4, 26), der mit dem Hostobjekt verbunden ist, um eine Objektdynamik des Zielobjekts (2, 24) im Verhältnis zum Hostobjekt zu erfassen,
ein erstes Sicherheitseinrichtungs-Stellglied (5), das mit dem Hostobjekt verbunden
ist, um eine erste Sicherheitseinrichtung (6) zu aktiveren,
wobei das System dadurch gekennzeichnet ist, dass es ferner einen Pre-Crash-Algorithmus zur Vorausberechnung der Position des Zielobjekts
innerhalb eines Koordinatensystems (OXY) umfasst, das bestimmt ist durch eine Querachse
(OX), die sich der vordersten Kante entlang erstreckt, und durch die Längsachse (OY)
des Hostobjekts, basierend auf der Zielobjektdynamik,
wobei der Algorithmus die vorhergesagte Querposition (Xtp) des Zielobjekts vom Ursprung zu einer vorhergesagten Zeit tp vorausberechnet, zu
der das Zielobjekt die Querachse (OX) kreuzt,
ein Sicherheitseinrichtungs-Steuergerät (9), das mit dem Hostobjekt verbunden ist, um basierend auf der vorhergesagten Querposition (Xtp) des Zielobjekts im Verhältnis zu einem ersten Schwellenbruchteil der Hostobjektbreite eine Kollisionspotenzialbeurteilung zu erstellen, und um das erste Sicherheitseinrichtungs-Stellglied in Reaktion auf die Kollisionspotenzialbeurteilung zu steuern.
Scharfschalteinrichtung für einen Airbag und Gurtstraffer mit Motorantrieb.
Erfassen eines Zielobjekts (2, 24) in einer Nahzone des Hostobjekts (3, 22),
Verfolgen einer aktuellen Objektdynamik des Zielobjekts (2, 24) im Verhältnis zu einem Koordinatensystem, das bestimmt ist durch eine Querachse, die sich der vordersten Kante entlang erstreckt, und durch die Längsachse des Hostobjekts,
Vorausberechnung der vorhergesagten Querposition des Zielobjekts vom Ursprung zu einer vorhergesagten Zeit, zu der das Zielobjekt die Querachse kreuzt, basierend auf der Zielobjektdynamik,
Erstellung einer Kollisionspotenzialbeurteilung durch einen Vergleich der vorhergesagten Querposition des Zielobjekts mit dem Schwellenbruchteil der Hostobjektbreite und
Betätigung einer Sicherheitseinrichtung in Reaktion auf die erstellte Kollisionspotenzialbeurteilung.
un capteur distant (4, 26) lequel est couplé à l'objet hôte afin de détecter une dynamique d'objet de l'objet cible (2, 24) par rapport à l'objet hôte,
un premier actionneur (5) de dispositif de sécurité lequel est couplé à l'objet hôte afin d'activer un premier dispositif de sécurité (6),
le système étant caractérisé en ce qu'il comprend en outre :
un algorithme d'avant collision lequel prédit la position de l'objet cible dans les limites d'un système de coordonnées (OXY) défini par un axe transversal (OX) qui s'étend le long du bord situé le plus en avant et par l'axe longitudinal (OY) de l'objet hôte, en fonction de la dynamique de l'objet cible,
l'algorithme prédisant la position transversale prévue (Xtp) de l'objet cible à partir de l'origine à un instant prévu (tp) lorsque l'objet cible croise l'axe transversal (OX),
un contrôleur (9) de dispositif de sécurité lequel est couplé à l'objet hôte afin de déterminer une évaluation du potentiel de collision en fonction de la position transversale prévue (Xtp) de l'objet cible par rapport à une première fraction seuil de la largeur de l'objet hôte, et pour piloter le premier actionneur de dispositif de sécurité en réaction à l'évaluation du potentiel de collision.
détecter un objet cible (2, 24) dans une zone proche de l'objet hôte (3, 22),
suivre une dynamique d'objet actuelle de l'objet cible (2, 24) par rapport à un système de coordonnées défini par un axe transversal qui s'étend le long du bord situé le plus en avant et par l'axe longitudinal de l'objet hôte,
prédire la position transversale prévue de l'objet cible à partir de l'origine à un instant prévu lorsque l'objet cible croise l'axe transversal, en fonction de la dynamique de l'objet cible,
déterminer une évaluation du potentiel de collision grâce à la comparaison entre la position transversale prévue de l'objet cible et une fraction seuil de la largeur de l'objet hôte, et
actionner un dispositif de sécurité en réaction à l'évaluation du potentiel de collision qui a été déterminé.
REFERENCES CITED IN THE DESCRIPTION
This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.
Patent documents cited in the description